[sold out] SKY RAY 6 x Cree XM-L2 T6 3-Mode Cool White Light Flashlight - Black (4 x 18650) US$ 25,76 free shipping @ DX

http://www.ebay.de/itm/NIKOS-P70N02LDG-TO-252-/221240683541?pt=Bauteile&hash=item3382f8ec15

or can also an US-Member sells me one or two and send it to germany?

Our illustirous RMM has them for sale on his web site and it currently shows 26 in stock:
Here.

His shipping terms are here for outside the USA:
Here.

:bigsmile:

Thank you for this information!

just write him a PM about the shipping costs ( if i remember 15 USD for a Flashlight - a few bucks for small parts)

he is nice and answers fast! :wink:

Ah dang!

This might help you rebuild your switch
https://oshpark.com/shared_projects/mrk56N4w

Just need a pushbutton, solder on, then solder on the wires, well after you fix the metal rim

Slowly but surely you can get your SRK back together…just be gentle with it

Good luck good sir!

It no biggie that the light is dead. I was not planning on keeping it original anyway. I just hate it when thing break when they are not MEANT to break LOL

But thank you for your concern :-)

Ugh… now I kind of regret it I didn’t buy one at the original price :Sp

My first batch of 3 shold be at my house when I hey home from work! I grabbed a total of 9 from the US direct dx at $23… glad I pulled the trigger

Got home from trying to catch a surf perch and mine was waiting in the mailbox. This is my first SRK so I’ve nothing to compare it to, but I’m very impressed with the build quality. For that price I was expecting junk. The cool white tint was not as bad as I feared and best of all… NO BLINKY!!! I don’t know why, defective driver? Ha, that is one defect I wish happened more often. My only complaint is PWM noise on low. I hate driver screech…

Now I’ll enjoy the honeymoon phase while watching you experts lead the way in mods. dchomak’s thread looks very interesting!

Push & hold switch button when the light is on, et voila, your blinkies will appear! :bigsmile:

I wanted to take out all the variables, and see how many amps the star would take hooked directly to 4 cells in parallel. Here is a picture of the star with the 6-XM-L2s.

What I did is remove the star from the light and disconnected it from the driver. I had already upgraded the leads to the star to 16 gauge wire to handle the expected increase in current. I made a 4 cell pack by soldering a copper buss to the postive and negative terminals of the cells. The cells are the purple Samsung 2800 from the laptop pull posted by Siskiyou. I charged them up to 4.18V and connected everything together.
The cell pack

The circuit.

The meter on the left is reading the voltage of the pack. The yellow meter in the middle is measuring the temperature of the star. I placed the thermocouple right up next to one of the LEDs’. The meter to the right is going to measure the current to the star. Because this is a direct drive hookup, I knew that the LEDs would be subject to thermal runaway and wanted to monitor that. It is hard to see the star as in this pic it is edgewise, but you can see the orange clamp that is holding the thermocouple. Here is a picture of that connection. The little silver blob to the right of the LED and under the yellow tab is the thermocouple.

Now a very short video (19 seconds) of direct drive to the star. No driver.

Notice the current goes up and the temperature increases. I only let the star go to 50C as it is not sinked and was rising fast. Anyway about 10A. I believe the limiting factor is the copper traces on the star, probably some resistance there. Here is an angle shot of those traces.

OK, now onto the driver. First thing I did when I opened up this light was replace the original MOSFET with a 70N02. The 70N02 is a low resistance FET. Let’s see how this driver does with that replacement, comparing it to no driver at all, direct drive. In otherwords, will this driver hinder the maximum performance possible? The light is back together, here is a pic of that setup. Weakest link here is the alligator clip connections. The clips are NOS “Mueller” a vintage good quality clip and the banana jacks are also NOS of good quality. The leads are 16 gauge.

Now another very short video of the test.

Once again, about 10A
I think what this shows is that with the simple, and cheap, mods of changing the FET to the 70N02 and using heavier lead wires, this light can become all that it can be. Don’t forget the spring mod. This time when I put it back together I added some thermal grease to the threads of the threaded sink.

In conclusion, We got a GREAT deal on this light! There can be some improvements to the output of this light with a couple of cheap and easy mods. I believe the traces on the star to be the ultimate limiting factor. Personally, I would (did) some simple and cheap mods to improve the output. Total cost: $23.10 (DX US) plus about $1 for the 70N02 and the new lead wires. Of course the spring mods are free.

If someone could figure out how to add copper wires on top of the traces to the LEDs on the star, we could cut down on the resistance there…………

Nice testing. I’ve burned out the traces on three lights that had all the emitters mounted on a single MCPCB. I’m not sure why. Heat?

Good stuff dchomak, makes my wait a bit easyer, thanx! Can you please measure the Amps (original vs modded) in LOW mode?

Very nice testing dchomak!

I have no SRK, nor plans to buy one, but I like seeing mods happening, very good information/pictures/video's :-)

Could always try stripping the solder mask over the traces and run a bead of solder to thicken the traces, then kapton tape the snot out of the top of the star

Are they the typical hi>lo hidden strobe? How bad was the PWM on the lower modes?

Wow…very nice SRK, definitely needs good heatsinking, is the star bonded to the pill or just silicon grease?

I wish my SRK had a pill, all I have is a star that sits on the shelf that was machined out where the threads “should” be…
(once I get the driver up and running right, I might go ahead and do the fill of epoxy behind it to see if that helps…yeah makes it permanent…but this is my first “test” SRK, definitely won’t be my last)

That is a scary thought, yes probably to much current and the trace did an imitation of a fuse. We have to be careful of that.

Too late to get stock readings on this light, good thing I have another coming from DX China. If it is the same exact light I will get the readings. I want to know too!

Spacing is pretty good. I can’t detect any PWM on low. I can hear it.
BTW, the older we get, the longer it takes for our eyes to adjust to changes in light levels. Maybe at my age I have “long persistence” eyes.

Maybe not scrape the whole trace, and not every one. Perhaps just the ones that are a bottle neck and perhaps just an area near the LED. I have some small gauge solid copper wire I can use (telephone wire). If need be that wire can be pounded flat first, insulation and all.
Or maybe connect to the little blob of solder coming out from each LED during the reflow process. That is risky though, may disturb the electrical and or the thermal connection.

Dchomak, first of all I want to thank you for taking the time to post your test findings. However, I disagree with your findings. That 6x XM-L2 MCPCB will pull FAR more than 10A on four good cells. In fact, I would be very surprised to see less than 20A on four good cells with heavy wire to the MCPCB. Mine pulled 13.2A from a single 18650, with the stock skinny wires, over 10A from a single IMR 18350. I am guessing that with 18AWG wire it will do 16A+ from a single 18650.

Something in your test setup is adding a lot of resistance.

If it pops may end up scraping off the solder mask then adding solder to the traces on the MCPCB, if that doesn't work I will jumper the traces with wires.

RMM, is that current draw measured with direct drive, stock driver, or with the MOSFET swap?

I was very aware of how resistance in the circuit could impact the results, that is why I set it up so as to try and eliminate any possible points of resistance. As you know, it only takes one bad connection to cause a problem.

If that is the case, then it could be the alligator clips, banana jacks or even the cells themselves. Voltage at the pack dropped 0.2V after the 10A draw, so the internal resistance of the 4 pack would be .02 ohm or 20 milliohm. Times that by 4 and you get 80 milliohm per cell. Is that about to be expected for that cell?
I am pretty sure that XM-L2s will draw less than an XM-L with 18650s in direct drive. But you are right in expecting much more, so was I. That is why I immediately suspected the traces on the MCPCB.

The light is back together, I will try a test later with nothing but solder joints at ALL connections, even the meter. With the driver in place if the current goes up, then we know.

Actually, the current draw in the video is slightly higher WITH the driver than WITHOUT. That means something is amiss here. This is a case where if I am wrong about the MCPCB, it would be GOOD news. I would be glad to be wrong about that.

Direct drive to the MCPCB with the stock skinny wires and a single cell.